SEARCH

SEARCH BY CITATION

Keywords:

  • glycosaminoglycans;
  • heparin;
  • immune thrombocytopenia;
  • Platelet Factor 4;
  • thrombocytopenia;
  • thrombosis

Abstract

  1. Top of page
  2. Abstract
  3. Background
  4. In vitro and in vivo mechanistic studies
  5. Conclusion
  6. Disclosure of Conflict of Interests
  7. References

Summary.  Heparin-induced thrombocytopenia (HIT) is an iatrogenic disorder that occurs in a small subset of patients receiving heparin. Twenty-five per cent (or higher) of affected patients develop limb or life-threatening thrombosis. The effectiveness of therapy is incomplete and may be complicated by bleeding. HIT is caused by antibodies that recognize the platelet chemokine, Platelet Factor 4 (PF4), complexed to heparin or to cellular glycosaminoglycans (GAGs). However, antibodies with the same apparent specificity are found in many more patients without clinical disease and the reason why so few develop HIT is uncertain. We propose that HIT antibodies recognize cell surface PF4/GAG complexes on intravascular cells, including platelets and monocytes that are dynamic and mutable. Heparin removes cell surface-bound PF4 in most individuals, but removal is incomplete in those with high pre-exposure surface-bound PF4 levels. Such individuals retain critically localized cellular antigenic complexes at the time antibodies develop and are at risk to develop HIT. This article reviews the scientific basis for this model and its clinical implications.


Background

  1. Top of page
  2. Abstract
  3. Background
  4. In vitro and in vivo mechanistic studies
  5. Conclusion
  6. Disclosure of Conflict of Interests
  7. References

Heparin-induced thrombocytopenia (HIT) is an iatrogenic autoimmune disorder associated with a high risk of limb- and life-threatening thrombosis, especially in those exposed to unfractionated heparin (UFH), which contains high molecular weight glycosaminoglycans (GAGs). The salient clinical features were first described in the 1970s [1], and the immune basis was appreciated soon thereafter [2]. However, it was not until 1992 that the platelet chemokine, Platelet Factor 4 (PF4) complexed to heparin or other high molecular negatively-charged molecules, such as other GAGs, was identified as the inciting antigen and target of HIT antibodies [3]. The pathogenesis had been thought to involve circulating PF4/heparin immune complexes that bind to the FcγRIIA receptor on platelets and related Fc receptors on monocytes, endothelial cells and other vascular cells that promote platelet activation and generation of thrombin, setting up an explosive feed forward, prothrombotic loop [4].

Identification of PF4 as the target antigen led to the development of ELISA kits to detect HIT antibody, which are used to supplement functional assays, such as the serotonin-release assay, which in turn, vary in sensitivity and are not widely available [5,6]. It was hoped that antibody detection would facilitate rapid diagnosis in complex medical situations where co-existing causes of thrombocytopenia and thrombosis are common. However, subsequent studies have shown that up to 50% of patients exposed to UFH in settings of intense platelet activation, such as cardiopulmonary bypass surgery, already have and/or develop anti-PF4/heparin antibodies [4], but only 1%–3% of such patients would develop HIT.

The unanticipated prevalence of HIT antibodies not only has limited the diagnostic value of ELISA-based diagnosis, but it also opened the question of why only a subset of patients with such antibodies develops HIT. Recent studies show an incomplete correlation with IgG isotype and titer [6] and no correlation with FcγRIIA polymorphisms [7]. The contribution of antigenic site specificity [4] has not been explored. The rapid onset of HIT, as early as day 5 after initial UFH exposure, is unusual for an IgG-mediated primary immune response [4]. Other patients develop thrombosis 1–2 weeks after the last identified exposure to UFH, which has been called ‘delayed HIT’, and its mechanistic basis remains unclear.

In vitro and in vivo mechanistic studies

  1. Top of page
  2. Abstract
  3. Background
  4. In vitro and in vivo mechanistic studies
  5. Conclusion
  6. Disclosure of Conflict of Interests
  7. References

PF4 and surface GAGs form HIT antigenic complexes

Several groups have reported that the molar ratio of UFH to PF4 modulates antigenicity. Binding of HIT antibodies is optimized at a 1:1 molar ratio of UFH and tetrameric PF4 [8,9]. We and others have shown that at this molar ratio, PF4 and UFH form high molecular weight complexes (>670 kDa), with each complex capable of binding as many as three antibody molecules [10]. An excess of either PF4 or UFH reduces both formation of these large complexes and antibody-binding.

Such antigenic complexes also develop on the platelet surface in the presence of free PF4. When PF4 is added to human or mouse platelets at increasing concentrations, binding of the murine monoclonal HIT-like antibody, KKO, [11] and activation of platelets by human HIT antibodies, follow a bell-shaped curve. Using human or mouse platelets, binding of human antibody and KKO peak at 50 μg mL−1 PF4 [12].

Platelets differ from other vascular cells in that the surface GAGs are composed almost exclusively of proteoglycans with chondroitin sulfate (CS) side chains [13]. PF4 binds to heparin with higher affinity than it does to cell surface GAGs, especially CS [14]. The effect of heparin on the development of platelet antigenicity depends on the amount of PF4 present. At low levels of added PF4 (e.g. 12 μg mL−1), even low concentrations of UFH markedly reduce antibody-binding. In the presence of concentrations of PF4 optimal for antigenicity (50 μg mL−1), higher UFH concentrations are required to disrupt antibody-binding. At yet higher concentrations of PF4 (e.g. 200 μg mL−1), antigenicity actually increases when UFH is added at concentrations that span the therapeutic range (1–4 μg mL−1), and then falls as the UFH concentration is increased further. The initial increase in antibody-binding may be attributable to the UFH removing PF4 from the platelet surface until the optimal PF4/GAG ratio is reached at which point large antigenic complexes predominate; antibody-binding falls once these complexes are disrupted by removal of additional PF4.

Monocytes have the potential to generate thrombin when activated, including by HIT antibodies [8,15,16] and might thereby contribute to thrombosis. Their cell surface GAGs contain CS, but also dermatan sulfate and heparan sulfates, [17] both of which bind PF4 with higher affinity than CS [14]. Monocytes synthesize hypersulfated GAGs when activated [17], which could potentially amplify their affinity for PF4. We compared the binding of HIT antibodies to monocytes and platelets. Binding of KKO to both cell types followed a bell-shaped curve with a similar optimal concentration of added PF4. However, monocytes bound significant amounts of HIT antibody at concentrations of PF4 where there was little binding to platelets. In addition, 2- to 4-fold higher concentrations of UFH were required to reduce antibody-binding to monocytes than to platelets at all concentrations of PF4 tested. Moreover, when monocytes were activated by addition of lipopolysaccharide, an increase in GAG sulfation was seen, which was associated with an additional 2-fold increase in the concentration of UFH needed to reduce surface antigenicity.

These studies suggest that HIT antibodies can recognize cell surface PF4/GAG complexes. Individuals who retain a sufficient density of antigenic complexes on platelets and other hematopoietic cell surfaces at the time of antibody formation are at risk to develop HIT, reflecting an interaction between persistence of antigen and antibody titer. Monocytes may be more susceptible to antibody-mediated activation than platelets at low concentrations of UFH, for example, when used as prophylaxis or after UFH has been stopped. Dynamic changes in antigen exposure reflecting heparin levels, extent of platelet activation and PF4 release, and GAG composition including extent of sulfation, may help to explain why most patients with HIT do not develop clinical disease and why patients are at risk to develop ‘delayed HIT’ even when platelet counts have returned to normal.

However so far, this proposed mechanism is based on in vitro studies. Are there in vivo data to support this model?

In vivo studies of the role of surface antigenic complexes in HIT

A number of years ago, a murine model of HIT was described involving the administration of KKO and UFH to mice that were double-transgenic for the expression of human PF4 and human FcγRIIA in platelets [18]. Subsequently, three transgenic lines of mice with 0.5×, 2–3× and 6× mean levels of human platelet PF4 were placed on the FcγRIIA transgenic background. Notably, unactivated platelets express surface human PF4 proportional to its content, perhaps as a result of leakage during proplatelet formation and impaired storage in the α-granules relative to mouse PF4. However, this observation provided us with a good model to study the human situation in which varying amounts of PF4 are expressed on platelets as a result of their activation in vivo because of underlying vascular disease and/or other medical conditions.

We would anticipate that these double-transgenic mice would not need UFH to develop thrombocytopenia after infusion of KKO based on our model because the antigenic PF4/GAG complexes would already be present on the platelets. This hypothesis was affirmed. While only mice transgenic for both human PF4 and FcγRIIA developed thrombocytopenia, the severity of the thrombocytopenia correlated with surface expression of human PF4 in the various transgenic lines [12]. Double-transgenic mice injected with KKO, but UFH, also developed a prothrombotic state evidenced in the rose bengal photochemical carotid artery injury model.

To further test our proposed model, we asked whether UFH or protamine sulfate (PS) would prevent KKO-induced thrombocytopenia. We hypothesized that high concentrations of UFH would disrupt and elute surface PF4/GAG antigenic complexes, as described in vitro, as would PS, which would compete with PF4 for binding to GAGs. A therapeutic dose of subcutaneous UFH (20 U/20 gm mouse) had no effect on the thrombocytopenia induced by KKO in double-transgenic mice expressing either the mid- or high-levels of PF4. Repeated, intravenous UFH prevented thrombocytopenia from developing in the mid-level mice, though not in the high-expressing mice, presumably because PF4 elution in the latter was offset by redistribution of residual PF4 into antigenic complexes. PS at concentrations used to reverse UFH clinically (40 μg/20 gm mouse) prevented antibody-mediated thrombocytopenia in both the medium and high PF4 content, double-transgenic mice [12].

We also examined the role of monocytes in HIT by injecting double-transgenic mice with clodronate-ladened liposomes prior to KKO. These liposomes eliminated circulating CD115+-monocytes for up to 2 days. Clodronate liposome-treated mice developed significantly more severe thrombocytopenia than untreated mice or mice infused with control liposomes, but were no longer prothrombotic in the photochemical carotid artery injury model. A similar outcome was seen when mice were treated with gadolinium chloride (50 mg kg−1), which inactivates monocytes [19]. These studies suggest that when monocytes are depleted and unable to remove antibody-coated platelets, more FcγRIIA-activated platelets circulate, where they can participate in the formation of intravascular microthrombi and intravascular platelet consumption. However, the arterial clots formed in monocyte-depleted animals were less efficiently formed or less stable than in controls. The effect of activated monocytes in the development of arterial and venous thrombosis and thrombocytopenia requires additional study.

Conclusion

  1. Top of page
  2. Abstract
  3. Background
  4. In vitro and in vivo mechanistic studies
  5. Conclusion
  6. Disclosure of Conflict of Interests
  7. References

Our current model of HIT is shown in Fig. 1. The clinical course of most individuals exposed to UFH is characterized by the scenario shown in Fig. 1A: platelets (top) have low to normal internal stores of PF4 and do not undergo intense chronic activation or release large amounts of PF4, which would bind to surface GAGs (right). When exposure to UFH occurs, the small amount of surface PF4 is removed to form circulating PF4/heparin complexes (bottom). Soluble PF4/heparin complexes stimulate antibody formation, perhaps based on their capacity to form colloidal particles carrying a net positive charge [20] (bottom). Circulating HIT antibodies may then bind to the soluble PF4/heparin (bottom), but the resulting concentration of immune complexes is below the threshold required to cross-link cell surface Fc receptors and activate platelets or monocytes. More importantly, there are no surface HIT antigenic complexes on platelets or monocytes for these HIT antibodies to bind with and to then activate these cells. The platelets remain quiescent, retain PF4 in their α-granules (left), and the patient is not susceptible to develop HIT (middle).

image

Figure 1.  Model of the role of platelet and monocyte surface PF4/GAG complexes on the pathogenesis of HIT. See text for a description of figures.

Download figure to PowerPoint

A smaller number of patients have greater stores of PF4 in their platelets (top, Fig. 1B). When these platelets are exposed to sufficient chronic intravascular activation, significant amounts of PF4 are released and reattach to cell surface GAGs (right). Upon heparinization, not all of the cell surface PF4 is removed, leaving potential PF4/GAG antigenic complexes on platelets and monocytes (bottom). Again, soluble PF4/heparin results in the development of HIT antibody development and perhaps antibody formation is intensified by the greater amount of circulating PF4/heparin complexes (bottom). While the presence of higher titers of HIT-like antibodies is contributory, our model suggests that the fact that more PF4/GAG antigenic targets remain on vascular cell surfaces at therapeutic concentrations of heparin is of greater importance. Once antibody is bound to the platelet surface and activates these platelets through their FcγIIA receptors, more PF4 is released, setting up a positive feedback loop of hematopoietic and likely vascular cell sensitization and activation (left). GAGs on activated monocytes become over-sulfated and bind PF4 with greater affinity (left). Antibody-coated, activated platelets are removed in the spleen, but are predisposed to adhere to each other, to other activated hematopoietic cells as well as to the vasculature, and become incorporated in nascent thrombi if removal from the circulation is impeded. Activated monocytes express tissue factor and release interleukin-8, which may contribute to the development of intravascular thrombosis, leading to clinical HIT (middle).

The distinguishing features of our model help to explain the seemingly paradoxical occurrence of thrombocytopenia and thrombosis, and why the procoagulant state persists after heparin has been discontinued and thrombocytopenia has resolved. Our model also suggests that steps proximal to the generation of thrombin, including antigen formation on cell surfaces, and Fc receptor-mediated platelet and monocyte activation, upstream of thrombin formation may be more rational targets for intervention than administration of thrombin inhibitors.

Most of our studies with the HIT-like monoclonal antibody KKO have been confirmed with HIT sera. However, we have not performed comparable studies of sera from individuals with high antibody titers but without clinical HIT to assess antigen specificity or other attributes of the antibodies themselves. We also realize that the model is predicated to a considerable extent on results obtained in vitro or in a murine HIT model and each aspect requires confirmation with patient samples. Finding that patients with high platelet PF4 content, surface PF4 expression and/or surface HIT antigenic complex are at greater risk to develop HIT, would permit identification of patients who require closer monitoring or drug avoidance. Detection of surface PF4 on platelets and/or monocytes in a heparinized patient with a positive HIT ELISA may allow a clearer diagnosis of HIT in medically complicated patients.

Finally, our model implies that high-molecular weight PF4/GAG HIT antigenic complexes are normally present on cell surfaces at sites of vascular injury and platelet plug formation. We have previously shown that platelet PF4 is necessary to optimize clot formation [21]. Either an excess or a deficit of PF4 impairs thrombus development. We hypothesize that the formation of PF4/GAG high molecular complexes underlie the observation that a specific level of PF4 inside of platelets is needed for thrombosis, consistent with the observation that high levels of PF4 is found in the α-granules of all mammalian species studied to date. The biologic mechanism by which PF4/GAG HIT antigenic complexes on cell surfaces optimize physiologic thrombus formation remains to be elucidated.

Disclosure of Conflict of Interests

  1. Top of page
  2. Abstract
  3. Background
  4. In vitro and in vivo mechanistic studies
  5. Conclusion
  6. Disclosure of Conflict of Interests
  7. References

The authors state that they have no conflicts of interest.

References

  1. Top of page
  2. Abstract
  3. Background
  4. In vitro and in vivo mechanistic studies
  5. Conclusion
  6. Disclosure of Conflict of Interests
  7. References
  • 1
    Rhodes GR, Dixon RH, Silver D. Heparin induced thrombocytopenia with thrombotic and hemorrhagic manifestations. Surg Gynecol Obstet 1973; 136: 40916.
  • 2
    Babcock RB, Dumper CW, Scharfman WB. Heparin-induced immune thrombocytopenia. N Engl J Med 1976; 295: 23741.
  • 3
    Amiral J, Bridey F, Dreyfus M, Vissoc AM, Fressinaud E, Wolf M, Meyer D. Platelet factor 4 complexed to heparin is the target for antibodies generated in heparin-induced thrombocytopenia. Thromb Haemost 1992; 68: 956.
  • 4
    Arepally G, Cines DB. Pathogenesis of heparin-induced thrombocytopenia and thrombosis. Autoimmun Rev 2002; 1: 12532.
  • 5
    Arepally G, Reynolds C, Tomaski A, Amiral J, Jawad A, Poncz M, Cines DB. Comparison of PF4/heparin ELISA assay with the 14C-serotonin release assay in the diagnosis of heparin-induced thrombocytopenia. Am J Clin Pathol 1995; 104: 64854.
  • 6
    Warkentin TE, Sheppard JI, Moore JC, Sigouin CS, Kelton JG. Quantitative interpretation of optical density measurements using PF4-dependent enzyme-immunoassays. J Thromb Haemost 2008; 6: 130412.
  • 7
    Arepally G, McKenzie SE, Jiang XM, Poncz M, Cines DB. Fc gamma RIIA H/R 131 polymorphism, subclass-specific IgG anti-heparin/platelet factor 4 antibodies and clinical course in patients with heparin-induced thrombocytopenia and thrombosis. Blood 1997; 89: 3705.
  • 8
    Cines DB, Rauova L, Arepally G, Reilly MP, McKenzie SE, Sachais BS, Poncz M. Heparin-induced thrombocytopenia: an autoimmune disorder regulated through dynamic autoantigen assembly/disassembly. J Clin Apher 2007; 22: 316.
  • 9
    Rauova L, Poncz M, McKenzie SE, Reilly MP, Arepally G, Weisel JW, Nagaswami C, Cines DB, Sachais BS. Ultralarge complexes of PF4 and heparin are central to the pathogenesis of heparin-induced thrombocytopenia. Blood 2005; 105: 1318.
  • 10
    Greinacher A, Gopinadhan M, Gunther JU, Omer-Adam MA, Strobel U, Warkentin TE, Papastavrou G, Weitschies W, Helm CA. Close approximation of two platelet factor 4 tetramers by charge neutralization forms the antigens recognized by HIT antibodies. Arterioscler Thromb Vasc Biol 2006; 26: 238693.
  • 11
    Arepally GM, Kamei S, Park KS, Kamei K, Li ZQ, Liu W, Siegel DL, Kisiel W, Cines DB, Poncz M. Characterization of a murine monoclonal antibody that mimics heparin-induced thrombocytopenia antibodies. Blood 2000; 95: 153340.
  • 12
    Rauova L, Zhai L, Kowalska MA, Arepally GM, Cines DB, Poncz M. Role of platelet surface PF4 antigenic complexes in heparin-induced thrombocytopenia pathogenesis: diagnostic and therapeutic implications. Blood 2006; 107: 234653.
  • 13
    Ward JV, Packham MA. Characterization of the sulfated glycosaminoglycan on the surface and in the storage granules of rabbit platelets. Biochim Biophys Acta 1979; 583: 196207.
  • 14
    Handin RI, Cohen HJ. Purification and binding properties of human platelet factor four. J Biol Chem 1976; 251: 427382.
  • 15
    Arepally GM, Mayer IM. Antibodies from patients with heparin-induced thrombocytopenia stimulate monocytic cells to express tissue factor and secrete interleukin-8. Blood 2001; 98: 12524.
  • 16
    Pouplard C, Iochmann S, Renard B, Herault O, Colombat P, Amiral J, Gruel Y. Induction of monocyte tissue factor expression by antibodies to heparin-platelet factor 4 complexes developed in heparin-induced thrombocytopenia. Blood 2001; 97: 33002.
  • 17
    Petricevich VL, Michelacci YM. Proteoglycans synthesized in vitro by nude and normal mouse peritoneal macrophages. Biochim Biophys Acta 1990; 1053: 13543.
  • 18
    Reilly MP, Taylor SM, Hartman NK, Arepally GM, Sachais BS, Cines DB, Poncz M, McKenzie SE. Heparin-induced thrombocytopenia/thrombosis in a transgenic mouse model requires human platelet factor 4 and platelet activation through FcgammaRIIA. Blood 2001; 98: 24427.
  • 19
    Van Schooten CJ, Shahbazi S, Groot E, Oortwijn BD, Van Den Berg HM, Denis CV, Lenting PJ. Macrophages contribute to the cellular uptake of von Willebrand factor and factor VIII in vivo. Blood 2008; 112: 170412.
  • 20
    Suvarna S, Espinasse B, Qi R, Lubica R, Poncz M, Cines DB, Wiesner MR, Arepally GM. Determinants of PF4/heparin immunogenicity. Blood 2007; 110: 425360.
  • 21
    Eslin DE, Zhang C, Samuels KJ, Rauova L, Zhai L, Niewiarowski S, Cines DB, Poncz M, Kowalska MA. Transgenic mice studies demonstrate a role for platelet factor 4 in thrombosis: dissociation between anticoagulant and antithrombotic effect of heparin. Blood 2004; 104: 317380.